CtrlK
BlogDocsLog inGet started
Tessl Logo

tessl-labs/intent-integrity-kit

Closing the intent-to-code chasm - specification-driven development with BDD verification chain

86

1.82x
Quality

92%

Does it follow best practices?

Impact

86%

1.82x

Average score across 14 eval scenarios

SecuritybySnyk

Advisory

Suggest reviewing before use

Overview
Quality
Evals
Security
Files

criteria.jsonevals/scenario-2/

{
  "context": "Tests whether the agent produces a complete set of planning artifacts from a feature spec: structured technical context, technology decisions with rationale, data model with state transitions, API contracts traceable to requirements, and spec quality assessment before planning.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "plan.md exists",
      "description": "A plan.md file is created at specs/001-notification-service/plan.md",
      "max_score": 5
    },
    {
      "name": "Technical Context fields",
      "description": "plan.md contains a Technical Context section with all required fields filled in: Language/Version, Primary Dependencies, Storage, Testing, Target Platform, Project Type, Performance Goals, Constraints, Scale/Scope — none left as empty or placeholder-only",
      "max_score": 12
    },
    {
      "name": "No bare Option labels",
      "description": "plan.md does NOT contain 'Option 1:', 'Option 2:', 'Option 3:' labels or '[REMOVE IF UNUSED]' markers from the template",
      "max_score": 5
    },
    {
      "name": "research.md with rationale",
      "description": "research.md exists and contains at least one technology decision with explicit rationale and at least one alternative considered",
      "max_score": 8
    },
    {
      "name": "data-model.md entities",
      "description": "data-model.md exists and defines at least two entities (e.g., Notification, Channel, RetryPolicy) with their fields listed",
      "max_score": 8
    },
    {
      "name": "State transitions in data-model",
      "description": "data-model.md includes the notification status state machine or state transitions (queued → sending → delivered/failed)",
      "max_score": 8
    },
    {
      "name": "contracts/ directory",
      "description": "At least one API contract file exists in specs/001-notification-service/contracts/",
      "max_score": 6
    },
    {
      "name": "Contracts reference spec requirements",
      "description": "The API contract(s) reference specific functional requirements by identifier (FR-XXX) or clearly trace to individual requirements from the spec",
      "max_score": 8
    },
    {
      "name": "quickstart.md exists with scenarios",
      "description": "quickstart.md exists and contains specific test scenarios or example commands for exercising the notification feature (not just generic instructions)",
      "max_score": 6
    },
    {
      "name": "No governance content in plan",
      "description": "plan.md does NOT contain project-wide principles, non-negotiable rules applying to all features, or team workflow/process requirements — governance content belongs in CONSTITUTION.md",
      "max_score": 8
    },
    {
      "name": "Spec quality assessment performed",
      "description": "The agent assessed the spec quality before planning: checked FR count, flagged any missing measurable criteria, identified unresolved clarifications, and reported a quality score or summary",
      "max_score": 12
    },
    {
      "name": "Plan decisions trace to spec FRs",
      "description": "plan.md or research.md references specific FR-XXX identifiers from the spec when justifying technical decisions (e.g., choosing a message queue to satisfy FR-003 about retry policies)",
      "max_score": 8
    },
    {
      "name": "context.json updated",
      "description": ".specify/context.json is updated with plan-related data (e.g., architecture node classifications, eval scores, or other dashboard metadata) without overwriting existing fields",
      "max_score": 6
    }
  ]
}

evals

README.md

tile.json